4 ms·

Many diseases are as simple as your eyes being brown vs blue - you simply have or lack particular proteins. I wish it were that simple. :( Eye color, for exam

Many diseases are as simple as your eyes being brown vs blue - you simply have or lack particular proteins. I wish it were that simple. :( Eye color, for example, is influenced by a combination of multiple genes. There’s no straightforward eye color gene. Likewise, it’s especially true for diseases when something like “stomach cancer” is actually a dozen different diseases where each is influenced by dozens of different genes that share symptoms. Combinatorial explosion is a real problem.

shloosh 9y ago
Yes, but for any disease there are a number of known gene polymorphisms that predispose you to the disease. These genes may interact with the rest of your genome in incredibly complex ways, but I bet all it would take to cure a disease in most cases would be to target the one or two most prevalent polymorphisms.
- aaavl2821 9y ago
  unfortunately that usually isnt the case. there are very few diseases where one mutation or molecule is truly causative. in many areas, weve approached the limits of how much we can treat disease by just targeting one molecule or mutation. this is an area where $50-70B of pharma R&D spend goes each year for the last few decades, and the yield of that research in terms of approved drugs is declining exponentially
- dekhn 9y ago
  "You bet"? This is one of the core questions in human genomics/health research today and nobody has been able to make a convincing argument in general that your proposed approach would work.
  - Nomentatus 9y ago
    Would you agree: The question being specifically whether in most cases necessary-even-if-not-sufficient genetic traits can be found ("genetic traits", as opposed to "genetic illnesses" since illness may not be the most likely result.) The problem being that although so far most of our exciting findings are of at-least-necessary genes (or plural genes, etc); obviously it's just these sort of cases that will produce high correlations and grab attention, so there should be a presumption of the equivalent of "publication bias." Necessarily, these will be the low-hanging fruit even if such cases are actually the exception, overall. Of course, mere and relatively small correlations between disease and genetic traits are extremely common; but there are lots of possible explanations of such findings. In some cases 16 percent of all genes have a small positive correlation to a given disease, and "if everybody is at fault nobody is at fault!"
- patall 9y ago
  While it may sound easy to change a polymorphism (and even there only for a single base, don't get me started on larger structural variants), do not forget that you have to apply this one single change to 10 ^ 14 cells (or even just 1 in 10000 cells is 10 ^ 10), while evading the immune system and not inducing any detrimental other mutations. For a software developer it always sounds so simple just to fix a single spelling mistake but that really isn't it. (+ obvoiusly the stuff from the other comments about how hard it is to determine the cause to begin with)
nycdatasci 9y ago
The original comment was not suggesting that all genetic diseases and cancer (!) can be easily cured. Based on similar techniques, many single-gene disorders are effectively cured. Unfortunately, our disease-specific drug approval process does not scale and prevents us from realizing the benefits of these techniques and technologies. As a personal example, a close family member was recently diagnosed with shwachman-diamond syndrome (SDS). I have confirmed with multiple leading physicians (Dana Farber, Mayo Clinic, etc) that all of the tools necessary to cure her are presently available. At a high level, this would involve: 1. Extracting marrow. 2. Isolating blood-cell producing stem cells & amplifying in a dish. 3. Modifying using a product like [A] 4. Validating/re-sequencing to confirm transduction. 5. Re-insert modified cells into patient. Yet doctors can't engage in this work without going through the lengthy drug trial process. So instead, they recommend much riskier treatment options including some that would complicate future gene-therapy attempts. [A] - https://crispr.sigmainformatics.com/templates/CRISPR%20Custom%20Editable%20PDF_v8_savable.pdf https://crispr.sigmainformatics.com/templates/CRISPR%20Custo...
- Brakenshire 9y ago
  > Unfortunately, our disease-specific drug approval process does not scale and prevents us from realizing the benefits of these techniques and technologies. On the other hand, something which is exceptionally well known to work should also attract huge and immediate capital to do the proper tests. Even for rare diseases, a sure thing would be funded. I think the difficulty is more about procedural or cell based cures vs. drug-based cures or treatments. The system is set up for the latter, but is not comfortable with the former.
  - jlg23 9y ago
    > Even for rare diseases, a sure thing would be funded. Only if there is a ROI. Which also means that even treatments for not-so-rare diseases don't get funded if the majority of patients won't be able to afford the treatment at a price that provides that ROI. Clinical trials in the US or Europe are insanely expensive and you cannot even budget them: Adverse reactions by "male patients over 45, born on a new moon Tuesday"? In worst case you restart the trial, excluding the problematic group to at least get the proof of safety and effectiveness for some part of the population so you can finally make some money...
Nomentatus 9y ago
True, and it's probably also worth noting that even "single gene" diseases can still be environmental illnesses; dependent on a modern condition or conditions that didn't occur a hundred thousand years ago or more; and nonexistent (despite the presence of that gene) back then.

shloosh 9y ago

Yes, but for any disease there are a number of known gene polymorphisms that predispose you to the disease. These genes may interact with the rest of your genome in incredibly complex ways, but I bet all it would take to cure a disease in most cases would be to target the one or two most prevalent polymorphisms.

aaavl2821 9y ago
unfortunately that usually isnt the case. there are very few diseases where one mutation or molecule is truly causative. in many areas, weve approached the limits of how much we can treat disease by just targeting one molecule or mutation. this is an area where $50-70B of pharma R&D spend goes each year for the last few decades, and the yield of that research in terms of approved drugs is declining exponentially
dekhn 9y ago
"You bet"? This is one of the core questions in human genomics/health research today and nobody has been able to make a convincing argument in general that your proposed approach would work.
- Nomentatus 9y ago
  Would you agree: The question being specifically whether in most cases necessary-even-if-not-sufficient genetic traits can be found ("genetic traits", as opposed to "genetic illnesses" since illness may not be the most likely result.) The problem being that although so far most of our exciting findings are of at-least-necessary genes (or plural genes, etc); obviously it's just these sort of cases that will produce high correlations and grab attention, so there should be a presumption of the equivalent of "publication bias." Necessarily, these will be the low-hanging fruit even if such cases are actually the exception, overall. Of course, mere and relatively small correlations between disease and genetic traits are extremely common; but there are lots of possible explanations of such findings. In some cases 16 percent of all genes have a small positive correlation to a given disease, and "if everybody is at fault nobody is at fault!"
patall 9y ago
While it may sound easy to change a polymorphism (and even there only for a single base, don't get me started on larger structural variants), do not forget that you have to apply this one single change to 10 ^ 14 cells (or even just 1 in 10000 cells is 10 ^ 10), while evading the immune system and not inducing any detrimental other mutations. For a software developer it always sounds so simple just to fix a single spelling mistake but that really isn't it. (+ obvoiusly the stuff from the other comments about how hard it is to determine the cause to begin with)

nycdatasci 9y ago

The original comment was not suggesting that all genetic diseases and cancer (!) can be easily cured. Based on similar techniques, many single-gene disorders are effectively cured. Unfortunately, our disease-specific drug approval process does not scale and prevents us from realizing the benefits of these techniques and technologies. As a personal example, a close family member was recently diagnosed with shwachman-diamond syndrome (SDS). I have confirmed with multiple leading physicians (Dana Farber, Mayo Clinic, etc) that all of the tools necessary to cure her are presently available. At a high level, this would involve: 1. Extracting marrow. 2. Isolating blood-cell producing stem cells & amplifying in a dish. 3. Modifying using a product like [A] 4. Validating/re-sequencing to confirm transduction. 5. Re-insert modified cells into patient. Yet doctors can't engage in this work without going through the lengthy drug trial process. So instead, they recommend much riskier treatment options including some that would complicate future gene-therapy attempts. [A] - https://crispr.sigmainformatics.com/templates/CRISPR%20Custom%20Editable%20PDF_v8_savable.pdf https://crispr.sigmainformatics.com/templates/CRISPR%20Custo...

Brakenshire 9y ago
> Unfortunately, our disease-specific drug approval process does not scale and prevents us from realizing the benefits of these techniques and technologies. On the other hand, something which is exceptionally well known to work should also attract huge and immediate capital to do the proper tests. Even for rare diseases, a sure thing would be funded. I think the difficulty is more about procedural or cell based cures vs. drug-based cures or treatments. The system is set up for the latter, but is not comfortable with the former.
- jlg23 9y ago
  > Even for rare diseases, a sure thing would be funded. Only if there is a ROI. Which also means that even treatments for not-so-rare diseases don't get funded if the majority of patients won't be able to afford the treatment at a price that provides that ROI. Clinical trials in the US or Europe are insanely expensive and you cannot even budget them: Adverse reactions by "male patients over 45, born on a new moon Tuesday"? In worst case you restart the trial, excluding the problematic group to at least get the proof of safety and effectiveness for some part of the population so you can finally make some money...

Nomentatus 9y ago

True, and it's probably also worth noting that even "single gene" diseases can still be environmental illnesses; dependent on a modern condition or conditions that didn't occur a hundred thousand years ago or more; and nonexistent (despite the presence of that gene) back then.